Syngap+/- CA1 Pyramidal Neurons Exhibit Upregulated Translation of Long MRNAs Associated with LTP

Abstract

In the Syngap^+/- model of SYNGAP1-related intellectual disability (SRID), excessive neuronal protein synthesis is linked to deficits in synaptic plasticity. Here, we use Translating Ribosome Affinity Purification and RNA-seq (TRAP-seq) to identify mistranslating mRNAs in Syngap^+/- CA1 pyramidal neurons that exhibit occluded long-term potentiation (LTP). We find the translation environment is significantly altered in a manner that is distinct from the Fmr1^-/y model of fragile X syndrome (FXS), another monogenic model of autism and intellectual disability. The Syngap^+/- translatome is enriched for regulators of DNA repair and mimics changes induced with chemical LTP (cLTP) in WT. This includes a striking upregulation in the translation of mRNAs with a longer-length (>2 kb) coding sequence (CDS). In contrast, long CDS transcripts are downregulated with induction of Gp1 metabotropic glutamate receptor-induced long-term depression (mGluR-LTD) in WT, and in the Fmr1^-/y model that exhibits occluded mGluR-LTD. Together, our results show the Syngap^+/- and Fmr1^-/y models mimic the translation environments of LTP and LTD, respectively, consistent with the saturation of plasticity states in each model. Moreover, we show that translation of >2 kb mRNAs is a defining feature of LTP that is oppositely regulated during LTD, revealing a novel mRNA signature of plasticity.

Keywords: Fragile X; LTD; LTP; Syngap; translation.

Figure 1.

The translating mRNA population in Syngap^+/− CA1 neurons is enriched for DNA repair proteins and is distinct from the population in Fmr1^−/y CA1 neurons. A, Schematic for TRAP-seq and total RNA-seq analysis of Syngap^+/− versus WT (N = 4 littermate pairs) from hippocampal CA1 neurons. B, Volcano plots for differential analysis of TRAP-seq data on the left and total RNA-seq on the right show substantial changes in translatome than the transcriptome of Syngap^+/−. Significant transcripts (adjusted p value < 0.1) being undertranslated or underexpressed are denoted in blue and overtranslated or overexpressed are denoted in red. C, Gene ontology (GO) analysis of transcripts upregulated and downregulated in translatome between Syngap^+/− versus WT. D, Gene set enrichment analysis (GSEA) of the Syngap^+/− (FDR < 0.05) shows a significant downregulation of glycoprotein metabolism, endoplasmic reticulum, and vacuole-related activities while upregulation of processes related to DNA repair, recombination, including ATP-dependent activity acting on DNA and transcriptional regulation. E, Schematic for TRAP-seq dataset comparison of Syngap^+/− versus WT to the TRAP-seq of Fmr1^−/y versus WT from hippocampal CA1 neurons. F, Quantification of transcripts shows 526 significant transcripts (p value < 0.01) are differentially translating in Syngap^+/− and only 25 of those overlap with the significant translatome in Fmr1^−/y (*p = 0.032). G, Transcripts significantly changed in Fmr1^−/y translatome are negatively correlated with Syngap^+/− translatome changes (r = −0.156, R² = 0.022, *p = 0.00048). H, Gene ontology (GO) analysis of transcripts significantly (p value < 0.01) altered in Syngap^+/ and Fmr1^−/y shows only few functional processes are regulated similarly in both mutants (p value < 0.05). I, To determine whether the gene sets altered in Syngap^+/− are similar to those altered in the Fmr1^−/y translating population, significantly changed gene sets (adjusted p value < 0.01) were compared with those significantly changed in the Fmr1^−/y population (adjusted p value < 0.01). This reveals a modest overlap of 25 gene sets (*p = 0.045) but 19 of these (85%) are regulated in opposite direction. J, The gene sets inversely modulated between Syngap^+/− and Fmr1^−/y regulate the chromatin remodeling, RNA localization and metabolism, transcription factor, and DNA-dependent activities among others, which are upregulated in Syngap^+/− while downregulated in Fmr1^−/y. Data supported by Extended Data Figures 1-1–1-3 and Tables 1-1–1-7.

Figure 2.

cLTP-specific translation changes in WT match basal changes in Syngap^+/− CA1 neurons, but diverge from changes in Fmr1^−/y CA1 neurons. A, Schematic of the TRAP strategy from wild-type (WT) hippocampal slices stimulated with 50 µM forskolin to induce robust cLTP chemical LTP (Chen et al.) followed by ribotag pulldown and RNA-seq on Camk2a-positive CA1 and CA3 neurons. B, LTP-specific significant transcripts (adjusted p value < 0.1) show small but significantly positive correlation with Syngap^+/− translatome (r = −0.096, R² = 008, *p = 0.00034) while notably negative correlation with Fmr1^−/y translatome changes (r = −0.224, R² = 0.049, *p = 8.95 × 10⁻¹⁵). C, Analysis of the LTP-specific significant transcripts in the Syngap^+/− translatome shows significant increase in LTP-upregulated transcripts but no change in LTP-downregulated transcript (Kruskal–Wallis test *p = 2.15 × 10⁻¹², post hoc two-sided Wilcoxon rank-sum test up *p = 2.96 × 10⁻¹³, down p = 0.055), while LTP-specific significant transcripts in the Fmr1^−/y translatome show significant opposing change in both groups—LTP-upregulated and LTP-downregulated transcripts (Kruskal–Wallis test *p < 2.2 × 10⁻¹⁶, post hoc two-sided Wilcoxon rank-sum test up *p < 2.2 × 10⁻¹⁶, down *p = 0.00078). Boxplots display the distribution of Log₂FoldChange values across LTP-upregulated and LTP-downregulated group of transcripts. The box represents the interquartile range (25th–75th percentile), the center line indicates the median, and whiskers extend to 1.5 times the interquartile range. Data beyond the whiskers are shown as outliers. D, Joint distribution analysis of LTP-specific transcripts between Syngap^+/− and Fmr1^−/y translatomes in a 2D density plot shows the positive distribution pattern of LTP-upregulated transcripts in Syngap^+/−. E, Analysis of the significantly upregulated LTP-specific transcript population that are also upregulated in Syngap^+/− translatome fraction identifies transcripts, which are involved in synaptic functions, intracellular signaling, neuronal morphogenesis, axonal transport, and cell adhesion. F, To determine whether the gene sets altered in Syngap^+/− are also altered in the cLTP translating population, significantly changed Syngap^+/− gene sets (p value < 0.01) were compared with those significantly changed in the cLTP population (adjusted p value < 0.01). This unveils an overlap of 35 gene sets (p = 0.054); nonetheless, majority of these (74.28%) are similarly upregulated in both. G, The gene sets that are alike and upregulated in both Syngap^+/− and cLTP are involved in dendrite morphogenesis, chromosome organization, transcription, and DNA-dependent regulatory activities among others. H, Comparison of the gene sets altered in cLTP (adjusted p value < 0.01) with the ones altered in Fmr1^−/y (p value < 0.01) shows a greater overlap of 164 (44.3%) terms (*p = 1.24 × 10⁻⁵⁷); however, 56% of these terms are changed in an opposite direction. I, Shared gene sets between cLTP and Fmr1^−/y regulate important processes such as mitochondrial function, different metabolic processes, and translation which is upregulated in Fmr1^−/y while downregulated with cLTP. The gene sets involved in axonogenesis, synaptic adhesion, postsynaptic specialization, and heterochromatin organization are upregulated with LTP but downregulated in Fmr1^−/y. Data supported by Extended Data Figures 2-1 and 2-2 and Tables 2-1–2-3.

Figure 3.

mGluR-LTD–specific translation changes in CA1 neurons match basal changes Fmr1^−/y but not Syngap^+/− hippocampus. A, Schematic of the TRAP strategy from wild-type (WT) hippocampal slices stimulated with a 5 min pulse of 50 µM S-DHPG that induces robust mGluR-LTD (Seo et al.) followed by TRAP-seq on CA1 pyramidal neurons. B, LTD-specific significant transcripts (adjusted p value < 0.1) show no correlation with Syngap^+/− translatome (r = −0.026, R² = −0.001, p = 0.533) while remarkably significant positive correlation with Fmr1^−/y translatome changes (r = 0.199, R² = 0.038 *p = 1.375 × 10⁻⁰⁶). C, Analysis of the LTD-specific significant transcripts in the Syngap^+/− translatome shows significant decrease in LTD- upregulated transcripts (Kruskal–Wallis test *p = 0.0001, post hoc two-sided Wilcoxon rank-sum test up *p = 2.74 × 10⁻⁰⁵, down p = 0.07). In contrast, LTD-specific significant transcripts in the Fmr1^−/y translatome show notably significant increase in LTD-upregulated transcripts (Kruskal–Wallis test *p = 2.66 × 10⁻⁰⁸, post hoc two-sided Wilcoxon rank-sum test up *p = 7.04 × 10⁻⁰⁹, down p = 0.057). Boxplots display the distribution of Log₂FoldChange values across LTD-upregulated and LTD-downregulated group of transcripts. The box represents the interquartile range (25th–75th percentile), the center line indicates the median, and whiskers extend to 1.5 times the interquartile range. Data beyond the whiskers are shown as outliers. D, Combined distribution analysis of LTD-specific transcripts between Syngap^+/− and Fmr1^−/y translatomes in a 2D density plot shows the positive distribution pattern of LTD-upregulated transcripts with Fmr1^−/y translation changes. E, Analysis of the significantly upregulated LTD-specific transcript population that are also upregulated in Fmr1^−/y translatome fraction identifies transcripts which are involved in apoptosis, ribosomal as well as mitochondrial functions, transcription regulation and cellular transport. F, Comparison of the gene sets significantly altered in Syngap^+/− (p value < 0.01) with the ones altered in LTD (adjusted p value < 0.01) reveals nonsignificant overlap of merely 4.7% (p = 0.3927527). G, To determine whether the gene sets altered in mGluR-LTD translating population match Fmr1^−/y translatome, significantly changed Fmr1^−/y gene sets (p value < 0.01) were compared with those significantly changed in the LTD population (adjusted p value < 0.01). This unveils a greater overlap of 21% with 78 terms (*p = 1.83 × 10⁻⁴⁶) and 91% of these terms are changed in similar direction. The gene sets that are upregulated alike in LTD and Fmr1^−/y are involved in ribosome and mitochondrial function, while similarly downregulated sets are related to neuronal activity and synaptic membrane functions. Data supported by Extended Data Figures 3-1 and 3-2 and Tables 3-1, 3-2.

Figure 4.

LTP and LTD induce distinct translatome shifts including opposite changes in synaptic transcripts. A, Schematic of the TRAP strategy from wild-type (WT) mouse hippocampal slices induced for mGluR-LTD (Sang et al.) and chemical LTP (Chen et al.). B, LTP and LTD comparison shows distinctly translated transcripts in both phenomena. Differential analysis was performed to identify synaptic plasticity related transcripts, i.e., LTP versus unstimulated control and LTD versus unstimulated control. LTP and LTD significant (adjusted p value < 0.1) transcripts were compared to find a common overlap and specific population of transcripts. C, Volcano plot of ribosome-bound translating population of transcripts in LTP and LTD. Significant transcripts (adjusted p value  < 0.1) being undertranslated are denoted in blue and overtranslated are denoted in red. Both LTP and LTD stimulation exhibit upregulated translation of immediate early genes such as Arc, Fos, Npas4, and Egr1 indicating neuronal activation. LTP-specific translation of Ppp1r15a, Btg2, and Nr4a among other transcripts is also remarkable. D, E, Gene ontology analysis of LTP- and LTD-specific transcripts shows their unique functions. LTP-specific transcripts predominantly regulate axonogenesis, cell projection, and dendrite development processes among others (D) while LTD-specific transcripts primarily regulate cytoplasmic translation, pre- and postsynaptic translation, metabolism, and energy precursors synthesis (E). F, GSEA analysis of the LTP versus unstimulated control TRAP-seq dataset identified over translation of gene sets related to synapse organization, cell–cell junction, and chromatin remodeling in LTP (adjusted p value < 0.1). The downregulated gene sets in LTP are involved in mitochondrial and ribosomal functions as well as cytoplasmic translation. G, GSEA analysis of the LTD versus unstimulated control TRAP-seq dataset identified over translation of gene sets related to ribosome, translation, and mitochondrial terms in LTD (adjusted p value < 0.1), while the downregulated gene sets are involved in pre- and postsynaptic membrane functions as well as cell–cell adhesion. H, To determine uniquely altered gene sets altered, a comparison of significant (adjusted p value < 0.1) gene sets identified by GSEA in both LTP and LTD TRAP-seq datasets was performed. This reveals a common pool of 101 gene sets (*p = 2.14 × 10⁻⁴⁶) and their remarkably opposite regulation between LTP versus LTD. The gene sets involved in synapse assembly, axon development, and regulation of synapse structure are upregulated in LTP but downregulated in LTD. The gene sets related to cytoplasmic and mitochondrial ribosome, ATP synthesis, and electron transport among other significant terms are downregulated in LTP but upregulated in LTD. Data supported by Extended Data Figure 4-1 and Tables 4-1–4-3.

Figure 5.

Translation of long (>2 kb) transcripts is bidirectionally altered by stimulation of LTP versus LTD, and this is mimicked in Syngap^+/− and Fmr1^−/y mutant CA1 neurons. A, Transcripts significantly changed in Syngap^+/− translatome (p value < 0.01) show significant positive correlation with longer CDS length (left, r = 0.25, R² = 0.06, *p = 4.56 × 10⁻⁰⁵). A binned analysis on CDS lengths of altered translatome shows that Syngap^+/− TRAP fraction exhibits upregulated translation of longer transcripts (two-sample z test; >4 kb vs all: z = 10.716, *p < 2.2 × 10⁻¹⁶, 2–4 kb vs all: z = 7.8933, *p = 2.94 × 10⁻¹⁵, 1–2 kb vs all: z = −0.92171, p = 0.35, <1 kb vs all: z = −7.6739, *p = 1.66 × 10⁻¹⁴). B, Transcripts significantly changed in Fmr1^−/y translatome (p value < 0.01) show significant negative correlation with longer CDS length (left, r = −0.24, R² = 0.057, *p = 2.32 × 10⁻⁰⁶). A binned analysis on CDS lengths of altered translatome shows that Fmr1^−/y TRAP fraction exhibits decreased translation of longer transcripts (two-sample z test; >4 kb vs all: z = −7.5046, *p = 6.16 × 10⁻¹⁴, 2–4 kb vs all: z = −10.079, *p < 2.2 × 10⁻¹⁶, 1–2 kb vs all: z = −1.5061, p = 0.132, <1 kb vs all: z = 12.301, *p < 2.2 × 10⁻¹⁶). C, Analysis of cLTP translatome (p value < 0.01) shows significant positive correlation with longer CDS length (left, r = 0.35, R² = 0.117, *p = 2.2 × 10⁻¹⁶). A binned analysis on CDS lengths of altered translatome shows that cLTP causes increased translation of longer transcripts (two-sample z test; >4 kb vs all: z = 13.987, *p < 2.2 × 10⁻¹⁶, 2–4 kb vs all: z = 10.401, *p < 2.2 × 10⁻¹⁶, 1–2 kb vs all: z = −4.9587, *p = 7.09 × 10⁻⁰⁷, <1 kb vs all: z = −13.1, *p < 2.2 × 10⁻¹⁶). D, Analysis of mGluR-LTD translatome (p value < 0.01) shows significant negative correlation with longer CDS length (left, r = −0.33, R² = 0.116, *p = 1.60 × 10⁻¹⁴). A binned analysis on CDS lengths of altered translatome in LTD exhibits decreased translation of longer transcripts (two-sample z test; >4 kb vs all: z = −14.638, *p < 2.2 × 10⁻¹⁶, 2–4 kb vs all: z = −13.203, *p < 2.2 × 10⁻¹⁶, 1–2 kb vs all: z = 1.2064, p = 0.22, < 1 kb vs all: z = 21.175, *p < 2.2 × 10⁻¹⁶). E, Gene ontology analysis of long transcripts (CDS length >2 kb) in LTP shows their functions in axon development, cell projection, and synapse structure organization (left). These transcripts are upregulated (log₂foldchanges, L2FC >0) and show largely similar alteration with Syngap^+/− while opposite patterns of alteration (L2FC <0) in LTD as well as Fmr1^−/y (right). F, Gene ontology analysis of long transcripts (CDS length >2 kb) in LTD shows they are related to axon development, cell projection, and synapse organization (left) similar to LTP but translation of these transcripts is decreased in LTD and Fmr1^−/y opposite to LTP and Syngap^+/− (right). G, H, Bidirectionally altered long transcripts in LTP and Syngap^+/− (G) versus LTD and Fmr1^−/y (H). Data supported by Extended Data Figures 5-1 and 5-2 and Tables 5-1–5-3.